Announcing the Final Examination of Peter Warren for the degree of Doctor of Philosophy
Non-destructive evaluation techniques (NDE) are critical for assessing the integrity, health, and mechanical properties of materials manufactured from various methods. These techniques generate vast amounts of data before, during, and after the manufacturing process. This large and always growing amount of data can be difficult to process and analyze using traditional methods. Machine learning (ML) offers a solution to this problem by providing powerful algorithms capable of learning patterns, identifying features, and finding relationships in large datasets. Various ML models are used in this work to make predictions, improve measurements and identify anomalies in data gathered through NDE techniques. In this work, neural networks are used to identify defects, classify defects, segment microstructure images, determine material health, and model manufacturing processes. Additive manufacturing enables the production of complex geometries and customized parts with reduced waste and lead time. The development of new material printing capability and techniques is necessary to expand its capabilities to produce high-performance parts with unique properties and functionality. A feasibility study of the implementation of Binder Jetting (BJT) is conducted on Martian and Lunar regolith in this work. The development of a novel data generation method on stainless steel manufactured via BJT to improve machine learning based microstructure measurements is also given in this work. The need for machine learning to process data gathered from NDE techniques is crucial to enhance the accuracy, efficiency and speed of analysis, which can lead to faster development and implementation of advanced manufacturing techniques.
Committee in Charge: Ranajay Ghosh, Jayanta Kapat, Seetha Raghavan, Gita Sukthankar
Read More